Google admits that it scrapes public data to train Bard

Right now, OpenAI, the company behind ChatGPT and DALL-E, is in the grips of a heated legal battle. The case says that the way that it obtains data to train ChatGPT- scraping public data off of the internet- is a massive violation of people’s privacy. It seems that Google is watching closely, as the company updated its privacy policy. In the process, it confirms that it scrapes data to train Google Bard.

We know that OpenAI is going to be fighting this legal battle for some time, and the outcome could have an impact on the course of AI going forward. We have an editorial on this case, and you can click here to read it.

Google confirms that it scrapes data to train Bard

Looking at the hot water that OpenAI’s in, it’s no surprise that Google is looking to cover its hide. The search giant is quite the collector of legal cases against it.

According to Gizmodo (via The Verge) Google updated its privacy policy on July 1st. The company didn’t really change much of the contents of the policy. However, it’s more clear about the fact that it scrapes data to train its AI products. This includes Bard, but it’s not limited to it. “…we use publicly available information to help train Google’s AI models and build products and features like Google Translate, Bard, and Cloud AI capabilities.”

The fact that it clarifies that it uses publicly available data to train its AI models, the company hopes, will keep it from catching fire. A major part of the case against OpenAI is the fact that it didn’t specify this. People are, unknowingly, having their data scraped to be used for AI.

Clarification isn’t the point

Google specified that it scrapes data, great; the company gets a gold star. But that’s not the point. This change just barely keeps the company within the gray area between legal and illegal. Maybe it will keep it out of the eyes of lawmakers and governmental agencies for now.

But this does nothing for the people it’s scraping data from. As AI takes more steps forward, it’s stepping on more and more people. Artists, journalists, creative writers, musicians, composers, developers, etc. Clarifying to these people that their data is being used to train the very AI models that will put them out of a job doesn’t make the situation better.

If Google, OpenAI, and other AI companies are really concerned about policing their creations, they’re going to have to do better than this. If not, they should clarify that they only care about money, and not the people affected by their products.